Annals of the Rheumatic Diseases 1993; 52: 182-184

Giant cells in arthritic synovium

Linda S Wilkinson, Andrew A Pitsillides, Jo C W Edwards

AbstractObjectives Giant cells are commonly

present in inflamed synovium, often inclose association with the intimal layer.The nature ofthese multinucleate cells hasbeen reassessed using new cytochemicaland immunochemical techniques.Methods Cryostat sections of non-

inflamed, rheumatoid arthritic andosteoarthritic synovia were analysed forthe presence of CD68 and non-specificesterase, markers associated with macro-

phages; activity of uridine diphospho-glucose dehydrogenase, associated withfibroblast-like synoviocytes; and tartrateresistant acid phosphatase and thevitronectin receptor subunit CD51,associated with osteoclasts.Results Giant cells were not seen in

non-inflamed tissue. In diseased tissuegiant cells in the intimal layer fell into twomajor groups: CD68 negative or dull cellswith high uridine diphosphoglucosedehydrogenase (UDPGD) activity sugges-tive of true synoviocyte polykaryons; andCD68 positive cells with low UDPGDactivity suggestive of macrophagepolykaryons. The two groups were seen insamples from patients with rheumatoidarthritis (RA) and patients withosteoarthritis (OA), but the former were

more prominent in OA and the latter inRA. Most CD68 positive giant cells alsoshowed tartrate resistant acid phospha-tase activity and prominent expression ofCD51. As such they were histochemicallyindistinguishable from osteoclasts, buttheir bone resorbing capacity remains

unknown.Conclusions Giant cells in arthritic

synovium appear to be of two types, one

related to true synoviocytes and one tomacrophages.

(Ann Rheum Dis 1993; 52: 182-184)

Department ofRheumatology,Faculty of ClinicalSciences,University CollegeLondon, London,United KingdomL S WilkinsonA A PitsillidesJ C W EdwardsCorrespondence to:Dr J C W Edwards,Arthur Stanley House,Tottenham Street,London WIP 9PG,United KingdomAccepted for publication29 October 1992

Giant cells are commonly present in diseasedsynovial tissue. Collins' and Grimley andSokoloff2 noted the presence of giant cells ofa particular morphology within or just beneaththe intimal layer in rheumatoid arthritic tissue.These cells are similar to mononuclear intimalcells in many respects and may carry processes

passing up to the intimal surface.Giant cells may derive from a range of

mononuclear precursors. Macrophage derivedgiant cells are commonly described3 and are

considered to be distinct from osteoclasts,4despite a number of common featuresincluding their probable origin from bonemarrow precursors.5 Fibroblast polykaryons

are rarely described in tissue, though they mayappear in fibroblast cell cultures.Developments in immunohistochemistry

and cytochemistry have made it possible todifferentiate individual synovial lining cellpopulations in tissue sections on the basis ofmetabolic activities or expression of surface orcytoplasmic markers.6 The aim of this studywas to analyse giant cell populations in syno-vium in detail with particular reference to twoquestions: the identity of the cells described byCollins' and Grimley and Sokoloff2 andevidence for the presence of cells showingfeatures commonly displayed by osteoclasts.

Materials and methodsNon-inflamed synovium was obtained fromthe limbs of five patients undergoingamputation for localised proximal sarcomata.Tissues were taken from macroscopicallynormal parts of the limb and showed nohistological abnormality. Inflamed synoviumwas obtained from 10 patients with osteo-arthritis (OA) and 10 with rheumatoid arthritis(RA) (definite or classical as defined byAmerican College of Rheumatology criteria)undergoing arthroplasty and was obtainedwithin 20 minutes of surgical removal.

Synovial tissues were dissected into 3-5 mmpieces which were snap frozen in n-hexane at-70°C. Sections were cut at -35°C on aBright's cryostat to a thickness of 7 pum and airdried. Sections for cytochemistry andcombined staining were reacted immediatelyfor enzyme activities. Sections forimmunohistochemical staining alone werestored at -70°C until required.

Uridine diphosphoglucose dehydrogenase(UDPGD) activity has been shown to be auseful marker for fibroblast-like synoviocytes intissue sections.6 The UDPGD activity wasassessed using the method of Mehdizadeh etal.7 Briefly, sections and cytospin preparationswere incubated at 37°C in a mediumcontaining UDP-glucose (3 mg/ml; Sigma)and NAD (0 3 mg/ml; Boehringer) in 30%(w/v) polyvinyl alcohol (Wacker) made up in0 05 M glycyl glycine buffer. The medium wassaturated with nitrogen, the pH adjusted to pH7-8 and, just before use, nitroblue tetrazolium(3 mg/ml) was added. In all instances thereaction time was 20 minutes. Controlincubations were performed using reactionmedium without UDP-glucose. After thereaction sections were washed in water, dried,and mounted in Aquamount (BDH), or, whenused for further staining, rinsed in phosphatebuffered saline (PBS).

Sections and cytospin preparations fordouble labelling were fixed in fresh acetone forfive minutes before reacting for enzymeactivity.

182

on February 24, 2022 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.52.3.182 on 1 March 1993. D

ownloaded from

Giant cells in arthritic synovium

Primary monoclonal antibimmunohistochemistry were Inising CD68, which is strongmacrophages),8 23C6 (recogni:of the vitronectin receptor, strby osteoclasts and weakly,macrophages),9 and 5B5 (rechydroxylase which is prese:synthesising cells such asendothelial cells) .o Sectionswith primary reagents for on(temperature. All subsequent irfor 30 minutes. Incubationsfluorescence were in PBS aprimary reagent being followeIgG fluorescein isothiocyanatcPBS containing 20% normalFor the alkaline phosphaphosphatase (APAAP) techidilutions were made in TRIS bpH 7-8. Primary monoclincubation was followed by rsimmunoglobulins, then Al(Dako), both made up in TRIS20% normal human serum. Simmunofluorescence were m(erol containing 1,4-diazabicy4(Sigma), and specimens stainin Aquamount.

Non-specific esterase activitmarker of macrophages

g r.I".-.0 :4, OP

%V. 0v ., 4#

# z-,V.!

* 9:

_.- 4K

is.It

,odies used for distinguishes these from fibroblastic cells inEBM 1I (recog- synovial lining.6 Normal human serum activity,ly expressed by was assessed using a standard ot naphthylsing CD5 1, part acetate method.1' When assessment wasongly expressed combined with assessment ofUDPGD activityif at all, by the UDPGD reaction was performed first.

zognising prolyl When combined with immunochemistry,nt in collagen assessment of normal human serum activityfibroblasts and was performed first.were incubated Acid phosphatase activity is characteristic ofe hour at room macrophages and osteoclasts. Acid phospha-ncubations were tase activity was shown using the method ofs for immuno- Barka,'2 with naphthol-AS-BI phosphate asit pH 7 6, the substrate. Tartrate resistant acid phosphatased by antimouse (TRAP) activity, which detects isoenzyme-5 ofe (Sera Lab) in acid phosphatase, present in large amounts inhuman serum. osteoclasts but not macrophages, was

tase-antialkaline visualised by inhibition of the acid phosphatasenique antibody reaction with 10 mM sodium tartrate.'3uffered saline at For double immunofluorescent staininglonal antibody preparations were incubated with one primaryabbit antimouse antibody, followed by a specific fluoresceinated'AAP (mouse) second layer. The tissue was then subjected to> buffered saline/ a blocking incubation with 20% normal mouseections used for serum before the addition of the secondounted in glyc- (biotinylated) primary antibody, visualisedclo [2.2.2] octane using a streptavidin conjugated reagent.ed with APAAP All double labelling procedures were

rigorously controlled by comparison withy is a traditional single labelled serial sections to ensure nowhich clearly interference between techniques. Assessment

ofUDPGD and TRAP activity was combinedwith anti-CD68, antiprolyl hydroxylase, andanti-CD51, visualised using a fluorescein

/ labelled second antibody.

4e.

4 *1..

/ ResultsGiant cells were not seen in non-inflamedtissue. They were identified in four of therheumatoid and five of the osteoarthritictissues examined. This is likely to be anunderestimate of their true frequency as giantcells were often only seen in a proportion ofsequential sections. When present theyoccurred at a rate of 1-1 1/m2 of section (up tofive in a X20 objective field).

Ninety per cent of giant cells lay eitheramong the deepest cells of the intimal layer orjust deep to the intima (fig 1). These cells fellinto two major groups: CD68 negative or dullcells with high UDPGD activity consistent

to the with true synoviocyte polykaryons and CD68timal positive, non-specific esterase positive cells

with relatively little UDPGD activity

...4f.: !4

... X Ct-

Figure 2 Rheumatoid arthritic synovium. (A) Reactedfor uridine diphosphoglucose dehydrogenase activity showing cells

ofhigh enzyme activity lining the tissue surface (uppermost) including a polykaryon (centre offield). (B) StainedforCD68 showing numerous positive cells close to the tissue surface, including many polykaryons (left of centre). Bars=50 ,um.

..

. W ,.Is&;,wFs:w *W.S ; ~~~~~~~~~~~~~~~~~~~,.

Figure 1 Rheumatoid arthritic synovium showing a giant cell (left of centre) closetissue surface (arrowed). In this instance the tissue is sectioned obliquely and the incells appear as a broad band interspersed with fibrin. Bar=25 Am.

183

:t..w:

u40AW.

i,-k

*.,:lt...'

AI

on February 24, 2022 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.52.3.182 on 1 March 1993. D

ownloaded from

Wilkinson, Ptsillides, Edwards

A

P..*E .

A'

_ f *

-

* ~ ~'

B1<

-, -

*4 -wS

C

Figure 3 Osteoarthritic synovium showing an area close tothe tissue surface containing a large number ofpolykaryonsstained (A) for non-specific esterase, (B) vitronectinreceptor, and (C) tartrate resistant acidphosphatase. Mostofthe polykaryons showing non-specific esterase activityalso show some stainingfor vitronectin receptor and tartrateresistant acid phosphatase. Bars=50 ,um.

consistent with macrophage polykaryons (fig2). Double labelling on single sections showedthese groups to be distinct. The groups wereseen in rheumatoid arthritic and osteoarthriticsamples but the former were more prominentin osteoarthritis and the latter in rheumatoidarthritis. CD68 positive giant cells containedup to 12 nuclei and had a 'shaggy' roughlycircular outline. CD68 negative giant cells hadup to four nuclei and had a smaller, smootheroval cell outline with, in some instances, longbroad cytoplasmic processes as seen inmononuclear synoviocytes.6The CD68 positive, non-specific esterase

positive giant cell group showed moderate orstrong acid phosphatase activity. Most of thesecells also showed TRAP activity and strongstaining with antibody 23C6 directed againstthe vitronectin receptor subunit CD51 (fig 3).There was a high degree of codistribution ofTRAP and vitronectin receptor staining.The few giant cells seen deep in the subin-

timal tissue were similar to the CD68 positivecells found in the vicinity of the intima.

DiscussionIt appears that Collins' and Grimley andSokoloff' were describing a mixed group ofgiant cells in rheumatoid synovium. Most havefeatures of macrophages, but a significantminority appear to be derived from fibroblast-like synoviocytes. In OA the latter grouppredominates. Moreover, there is evidence thatmany of the CD68 positive giant cells expressone or more markers which have beenassociated with osteoclasts.'4The presence of synoviocyte giant cells raises

the possibility that like other polykaryonforming cells the synoviocyte is an end stage

cell type which rarely, if ever, divides.Attempted mitosis in giant cells is thought tolead to cell death because of the lack of anymachinery for apportioning chromatin frommore than two sets of chromosomes correctlyto daughter cells. Polykaryons were not foundin normal tissue, however, and their presencein diseased tissue may represent an aberrantresponse to local factors such as cytokines'5incidental to the proliferative capacity ofsynoviocytes.From the practical point of view the finding

of two types of giant cell in the liningemphasises the importance of using doublelabelling to ensure correct identification of apotentially heterogeneous population of cellswhen studying properties such as enzymeactivities or expression of surface markers.The definition of osteoclasts is a matter of

continuing debate. Osteoclasts may be definedas giant cells which are resorbing bone or, byproxy, as cells carrying markers which havebeen associated with cells resorbing bone,including vitronectin receptors and TRAP.Conceptually, osteoclasts may also be definedas cells of a lineage which have the potential toresorb bone, whether or not this potential isrealised under the conditions of study. Thefindings of this study suggest that in theinflamed synovial intima giant cells displayingmacrophage features tend also to expressfeatures associated with osteoclasts. Until therelation between osteoclasts and macrophagederived polykaryons, in terms of lineage, isclearer the relation between expression ofhistochemical markers and the cells' ability todegrade bone remains in doubt.

1 Collins D H. The pathology of articular and spinal diseases.London: Edward Arnold, 1949.

2 Grimley P M, SokoloffL. Synovial giant cells in rheumatoidarthritis. AmJ Pathol 1966; 49: 931-54.

3 Chambers T J. Multinucleate giant cells.JPathol 1976; 126:125-48.

4 Vaes G. Cellular biology and biochemical mechanism ofbone resorption. Clin Orthop 1988; 231: 239-71.

5 Marks S C, Popoff S N. Bone cell biology: The regulationof development, structure and function in the skeleton.AmJlAnat 1988; 183: 1-44.

6 Wilkinson L S, Pitsillides A A, Worral J G, Edwards J C W.Light microscope characterisation of the fibroblast likesynovial lining cell. Arthritis Rheum. 1992;35:1179-84.

7 Mehdizadeh S, Bitensky L, Chayen J. The assay of uridinediphosphoglucose dehydrogenase activity: discriminationfrom xanthine oxidase activity. Cell Biochem Funct 1991;9:103-10.

8 Franklin W A, Mason D Y, Pulford K, et al. Immuno-histological analysis of human mononuclear phagocytesand dendritic cells by using monoclonal antibodies. LabInvest 1986; 54: 322-26.

9 Horton M A, Lewis D, McNulty K, Pringle J A S,Chambers T J. Monoclonal antibodies to osteoclastomas:definition of osteoclast-specific antigens. Cancer Res 1985;45: 5663-9.

10 Hoyhtya M, Myllyla R, Siuva J, KivirkkoK L, Tryggvason K.Monoclonal antibodies to human prolyl-4-hydroxylase.EurJBiochem 1984; 141: 477-82.

11 Gomori G. Histochemistry ofhuman esterases. J HistochemCytochem 1955; 3: 479-84.

12 Barka T. A simple azo dye method for histochemicaldemonstration of acid phosphatase. Nature 1960; 187:248-50.

13 Yam L T, U C Y, Lam K M. Tartrate resistant acidphosphatase in leukaemic reticuloendotheliosis. N EnglJ7Med 1971; 284: 357-60.

14 Athanasou N A, Alvarez J I, Ross F P, Quinn J M,Teitelbaum S L. Species differences in the immuno-phenotype of osteoclasts and mononuclear phagocytes.Calcif Tissue Int 1992; 50: 427-32.

15 Enelow R I, Sullivan G W, Carper H T, Mandell G L.Induction of multinucleated giant cell formation from invitro culture ofhuman monocytes with interleukin-3 andinterferon gamma: comparison with other stimulatingfactors. AmJ Respir Cell Mol Biol 1992; 6: 57-62.

184

*n

1

on February 24, 2022 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.52.3.182 on 1 March 1993. D

ownloaded from